Should software for clinical users be different from software for administrative users such as travel agents? Givien the completely different context the answer seems to be obviously yes. Still, clinical users have been confronted with the same kind of software as administrative users for decades – and have complained about that for decades.
For two decades business processes have been the guiding metaphor for clinical software. In many cases commercial software vendors have used their standard software packages and tried to refurbish it for clinical applications. Dissatisfaction with such software is abundant. Using the computer is associated with wasting time rather than with gaining added value for patient care.
In this research project we want to find out why should and how could clinical software be made such that it truly supports clinical work. Our hypothesis is that in addition to the institutions' processes the software should reflect the values of the users. Thus, we focus on value-based requirements engineering.
Value-based requirements engineering is a new approach within software engineering. By its inception, it tries to capture values from the management and the user perspective and to profit from them while specifying software requirements. If values are known, they provide a deeply rooted understanding of management goals and user needs, and thus explicit rationale for decision making and prioritization. Therefore, value-based requirements engineering seems an opportunity to foster both, better software and better requirements engineering processes. In this project we focus on the users' values. The research goal is to gain insight in users' values and from that to develop a method to reflect users' values during requirements engineering.
The context of our project is the field of clinical software. Medical care differs from other industries in a number of aspects some of which are important for this project. The two most important aspects are that patient cures are much less predictable and more complex than processes in other industries, and that doctors are highly skilled, highly intrinsically motivated and autonomous actors, rather than mechanical users of tools. In this setting the meaning of "users' values" has to be extended. We start from the hypothesis that actual values in individuals – the doctors and somewhat differently also nurses – outperform official values of the hospital as to their influence on what software is required.
Therefore, our research agenda is to elicit the values of doctors and nurses, their utilization of existing software and their degree of satisfaction respectively dissatisfaction with that software, and to develop a theory that links individual values to software requirements. Furthermore, we want to extend current requirements engineering methods to reflect this theory. This should enable requirements engineers to capture software requirements reflecting both institutional processes and users' values.
To achieve this tight linkage between values and requirements, methods from the social sciences have to be related to typical methods in software engineering. Interviews, questionnaires and recordings of utilization scenarios for selected types of clinical software are used as an empirical foundation of the actual state of users' values and how they relate to software properties. Approved methods from goal- and scenario-based requirements engineering are extended to include such value related information. We aim at instruments that are suitable for use in software development to capture and model values, requirements, and how they are related to each other. For the medical domain we want to get additional insights into prevailing value inventories which are typical for clinical users.
